Wages and Discrimination

Via Tyler Cowen, what looks to be some pretty important new research -- "Prejudice and the Economics of Discrimination" by Kerwin Kofi Charles and Jonathan Guryan of the University of Chicago. Here's the abstract:

This paper tests the predictions about the relationship between racial prejudice and racial wage gaps from Becker's (1957) seminal work on employer discrimination - something which has not previously been done in the large economics discrimination literature. Using rich data on racial prejudice from the General Social Survey, we find strong support for all of the key predictions from Becker about the relationship between prejudice and racial wage gaps. In particular, we show that, relative to white wages, black wages: (a) vary negatively with a measure of the prejudice of the "marginal" white in a state; (b) vary negatively with the prejudice in the lower tail of the prejudice distribution, but are unaffected by the prejudice of the most prejudiced persons in a state; and (c) vary negatively with the fraction of a state that is black. We show that these results are robust to a variety of extensions, including directly controlling for racial skill quality differences and instrumental variables estimates. We present some initial evidence to show that racial wage gaps are larger the more racially integrated is a state’s workforce, also as Becker's model predicts. The paper also briefly discusses familiar criticisms and extensions of the standard Becker model, including an argument of our own which, like some recent work, shows that the model's main predictions can be shown theoretically to survive the effects of long run competition.

On the other hand, seven of America's leading phrenologists assure me that wage gaps are due to inherent genetic inferiority and anyone who says otherwise is a creationist.

Comments

Charles and Gufyan must believe in creationism. William Saletan says so.

Posted by Tom Hilton | December 27, 2007 10:59 AM

Many of the opponents Yglesias satirizes--Rushton, Sailer, Lynn--advocate policies inimical to the interests of members of some one ill-defined "race." Saletan was, as he has more or less acknowledged, foolish to buy into all their research as if they weren't possibly motivated by racial animus to reach the conclusions they did.

Nevertheless, the structure of the argument does allow Yglesias' abrupt dismissal. Anyone with eyes can see that racial animus and discrimination pervade society still today, and together with the historical momentum of even greater animus and discrimination in the past, they afford powerful explanations of any black-white gaps observed today. For all practical purposes, reductions in this animus and discrimination should be the focus of policy (along with non-race-based progressive measures whose beneficiaries will often be black).

But the hypothesis is not that environmental factors are of negligible explanatory import, nor that genetic factors operate independently of and apart from environmental factors. What's hypothesized is interaction--that genetic factors play a causal role, but only in and through interaction with an array of environmental factors. The hypothesis is that, so far as population averages are concerned, the effect of genetic factors is (significantly) greater than zero. Just as lactose-tolerance was selected for by pressures not uniform across interbreeding populations, so also was some characteristic like "g" or whatever it is that underlies performance on intercorrelated standardized tests like the IQ tests and the SAT.
The hypothesis is that over hundreds of generations, thousands of years, genetic characteristics spread within one population and not another--not much controversy there--and that among these characteristics is a sort of genetic underlay of "g" or something of that sort--lots of controversy here. Also lots of controversy surrounding even the finding that genetically distinct populations conform to geographic groupings. But unless Science's acceptance policies are badly flawed, he controversy can't be resolved by sarcastic dismissal.
http://www.sciencedaily.com/releases/2005/01/050128221025.htm
http://query.nytimes.com/gst/fullpage.html?res=9500E3DE1E3DF933A15751C1A9649C8B63
Now suppose that the notion of a difference in genetic underpinnings of some g-like factor is mere piffle. Well then, it will soon--in a decade or two--have been demonstrated to have been so, and Saletan's wounds will never be salved. For with the mapping of the genome, every possible "genetic underpinning" of "g-like abilities" will have been observed--every genetic distinction between say the five Feldman racial/geographical groupings will have been observed. Perhaps no such distinctions will be found. Piffle proved. Perhaps whatever distinctions are observed look to have no mental correlates of any importance. Piffle proved.
What Saletan speculated was that in the event distinctions will be observed and they will look to be correlates (causes) of g-like mental ability. That is, "intelligence" will prove to be a set of performances based in the interaction of genes and environment, where the genes in question do vary across Feldman's five groupings.
I really don't know what will come of this ongoing research. I wouldn't be surprised if selection for anything "g-like" is uniform enough across environments that no genetic factors "underlying" "g" or something like it are there to be found. My only point is that the genetic hypothesis is not so readily to be dismissed as Yglesias suggests. It is an interaction hypothesis, and the undoubted importance of powerful environmental factors does not suffice to prove the non-existence of genetic differences. In a decade or two, the questions will be resolved. Then either Yglesias will be entirely vindicated--or he will be scrambling to find Saletan's suggestions as to how to accommodate the findings of racial difference.

Posted by Mugwump | December 27, 2007 12:18 PM

Good thing there is no racism. White folks tell me so.

Posted by sherifffruitfly | December 27, 2007 12:25 PM

...the model's main predictions can be shown theoretically to survive the effects of long run competition.

This is a key point (if it stands up to further scrutiny)

It flies in the face of the long-standing presumption of the competitive model that, somehow, market processes can be relied upon to remove the deleterious effects of discrimination (conceived to be an "irrational" non-market factor), in the absence of coordinated action to counteract discrimination.

But the core of Becker’s model rests upon the idea of discrimination arising from autonomous preferences of individuals. This also makes it a non-market factor. It is a conception that falls apart when one views the market itself as a dynamic process emerging out of a path-dependent history evolving through slavery, Reconstruction, Jim Crow, and corporate centralization.

Posted by Justin X | December 27, 2007 12:38 PM

To measure the voltage of a battery, you hook it up to a voltmeter. The result is unaffected by the degree of effort which the battery exerts.

Does this statement hold true if you substitute "intelligence" for "voltage" and "person" for "battery" and "IQ test" for "voltmeter"?

Put another way, is it a coincidence that the two groups put forward as having high test scores -- American Jews and Asians -- are groups in which (anecdotally at least) parents typically put pressure on their children to perform well academically?

I trot out these questions here and now because I do not assume that everyone working in this field is an idiot, so I expect that some attempt has been made to address them, and it sounds as though mugwump must know what the discipline has to say about them.

Posted by roac | December 27, 2007 12:52 PM

Obviously it is possible to hypothesize that there are some genetic differences between the "races" in regard to performance of certain specific mental tasks. But this hypothesis is very very far from being confirmed. This does not stop certain people form blathering on and on about this hypothesis as if it is fact. They occasionally will throw in a hedge here and there, but when they really get going you discover that for them the science is settled. And then you start investigating who these people are and you see that they say unambiguously racist things, that they have unfounded theories about penis-size correlating to intelligence, that their statistical work is controversial, and most interesting (to me), that they are usually conservative (having certain races be genetically inferior assuages to some degree white male guilt about their current privilege, a conservative sore spot.)

So I see no reason to why people keep speculating about this issue. The company is dubious the speculations are unwarranted and the speculating does great harm to our society. And I suspect the speculators motives.

Posted by cw | December 27, 2007 1:19 PM

"though mugwump must know what the discipline has to say about them..."

I would not be so quick to assume that mugwump knows anything. Perhaps the greatest insult to science is given by those who hope to find "in a decade or two" the "race/genetic/intelligence" trifecta but, in the meantime, plan on keeping the water as muddy as possible by suggesting the obvious as if it were some ground-shaking profundity.

"It is an interaction hypothesis, and the undoubted importance of powerful environmental factors does not suffice to prove the non-existence of genetic differences."

Is it now? Perhaps you can indicate what percentage of "intelligence" is determined by the environment and what percentage can not be explained by the environmental factors. Can you also let me know which genes are involved in shaping intelligence and also happen to cause superficial physiological differences?

There is no evidence, I repeat NO evidence that suggests that environmental factors cannot, in fact, explain ALL variances in intelligence. Nor is there any particularly good reason to believe otherwise. Now those who wish to ply this garbage science will spend a great deal of time discussing important "scientific" findings, interaction variables and factorial analysis, and genetic markers and genetic breakthroughs (though, not surprisingly, they tend to leave out the gene regulation stuff, since that reality tends to sour the simplistic theories that they like to put forward). But if you bother them long enough (and they are actually honest) they have no choice other than to agree with the first sentence of this paragraph (evidence to the contrary being non-existent).

Assuming a reasonable and open-minded demeanor while tut-tutting those who point out the vile nature of some of those who push this crap is yet another technique to cause confusion. "Only time will tell..." is especially pathetic, since with each day that passes, the neo-eugenicists arguments become more irrelevant as our understanding of not just genes, but gene regulation becomes more and more robust. And the more we learn, the more we understand that the first sentence of the previous paragraph remains true and the only reason to think otherwise is simple bias based not on science but upon willful ignorance.

Posted by Rihilism | December 27, 2007 1:42 PM

"I wouldn't be surprised if selection for anything "g-like" is uniform enough across environments that no genetic factors "underlying" "g" or something like it are there to be found."

I've often wondered about this myself. It isn't hard to understand the environmental pressures that make genetic tendencies favoring high melanin production adaptive in Zaire and low melanin production adaptive in Sweden, squat, fat bodies adaptive in the Yukon and tall, lean bodies adaptive in Kenya, and so on. But I've never figured out what environmental pressures would select against intelligence, or disproportionately favor intelligence in some environments rarther than others. Indeed, I suspect that, if anything, the harsh, unforgiving hunter-gatherer environments in which our darker brethern have to survive might select for intelligence while those of us descended from European proles crammed into filthy cities might be the result of selection for resistance to infectious disease rather than smarts.

Posted by CJColucci | December 27, 2007 3:52 PM

Estimates of heritability of any trait are less than straightforward, and those for intelligence less still. But one study comparing identical twins to fraternal found that the correlation the correlation between identical twins’ intelligences to be 0.86 compared to the correlation of .60 when comparing the intelligences of fraternal twins (Plomin & Spinath, 2004). A different study found the correlation of intelligence between identical twins raised in separate environments to be 0.72 (Toga & Thompson, 2005). Or, as Neisser et al summed up in their 1996 American Psychologist response to Murray and Herrnstein, “Across the ordinary range
of environments in modem Western societies, a sizable
part of the variation in intelligence test scores is associated
with genetic. differences among individuals. Quantitative
estimates yary from one study to another, because many
are based on small or selective samples. If one simply
combines all available correlations in a single analysis,
the heritability (h’) works out to about SO and the between-
family variance (c2) to about .25 (e.g., Chipuer,
Rovine, & Plomin, 1990; Loehlin, 1989).”

There is no evidence, I repeat NO evidence that suggests that environmental factors cannot, in fact, explain ALL variances in intelligence. Nor is there any particularly good reason to believe otherwise. is the norm, but I expect that Chelsea Clinton is pretty smart.

I think this statement is a little strong—I’m not sure that NONE of the evidence cited by Murray and Herrnstein has ANY validity. But as to the reason to believe otherwise, this seems to me to be the case:
1. Mutations spread rather quickly among populations separated from one another. Lactose tolerance, melanism and blue eyes are instances.
2. Since homo sapiens’ great migration from Africa commencing perhaps only 50 or 60,000 years ago, populations have been separated from each other so that differing mutations have appeared and spread through them, in response to differing social, climatic, environments.
3. There is evidence that brain differences are among those that have in fact appeared in some populations and not others. (“The gene Microcephalin (MCPH1) regulates brain size and has evolved under strong positive selection in the human evolutionary lineage.” http://www.sciencemag.org/cgi/content/abstract/309/5741/1717)
4. So maybe there’s no particularly good reason to believe that brain differences have arisen between populations that remained in Africa, those that settled in Australia, those that settled along the coastlines east of Africa, those that settled north of Africa, & etc. But there’s SOME reason—the reason of the nature, the SOP, of genetic differentiation of separated populations—to suppose that differences have arisen, or even that it would be surprising if NONE have.

Is this anything more than a straightforward application of the theory of natural selection?

Posted by Mugwump | December 27, 2007 5:22 PM

"But there’s SOME reason—the reason of the nature, the SOP, of genetic differentiation of separated populations—to suppose that differences have arisen, or even that it would be surprising if NONE have."

Why? Why do you think IQ is such fertile ground for selection? What leads you to make this assumption. The individual diffferences between the races are tiny. 99.99 something % of the human genome are identical across races. Nothing in this 99.99% has changed during the time the races were geologically seperate. Why do you think IQ ability would be part of the tiny chunk of genome that changed? Why aren't we speculating on liver function or toenail composition?

Posted by cw | December 27, 2007 5:59 PM

"Estimates of heritability of any trait are less than straightforward, and those for intelligence less still. But one study comparing identical twins to fraternal found that the correlation the correlation between identical twins’ intelligences to be 0.86 compared to the correlation of .60 when comparing the intelligences of fraternal twins (Plomin & Spinath, 2004)."

Just to clarify, my statement that there is "no evidence that environment cannot explain all differences" referred to the differences between "races". Often the argument is turned to imply that there is no genetic factor to intelligence, I make no such claims either for or against and have no reason to believe that genetics can't influence any apparently “heritable” trait. Exactly what the inherited IQ genes are that may or may not influence intelligence, I can't say (nor can anyone else). And, I've no opinion as to Chelsea's intelligence, but would imagine that (despite her father's dalliances) that she grew-up in a positive and nurturing environment.

I'd add that there is nothing in your points 1-4 to suggest that "race" has anything to do with variations in "IQ" in populations exposed to different environmental conditions. In fact natural selection makes it clear that animals with virtually identical genetic make-up can develop in a myriad of different ways due to environmental influences. The key, of course, is that potential is true for the species as a whole.

Posted by Rihilism | December 27, 2007 6:48 PM

"Put another way, is it a coincidence that the two groups put forward as having high test scores -- American Jews and Asians -- are groups in which (anecdotally at least) parents typically put pressure on their children to perform well academically?"

For Asians living in the United States, higher IQ tests than whites on average only occurred after financial success had become the norm for the average Asian families. In the 1970's, as Malcolm Gladwell pointed out, people starting saying that Asian children's IQ tests on average had surpassed white children's. However, they were comparing Asian children's scores on the 1950's version of the IQ test with white children's scores on the 1970's version of the IQ test, which was a harder test (the IQ test is made harder every generation due to the Flynn effect of a secular rise in IQ scores over time). As Gladwell notes, someone who was an Asian child in the 1970's who scored the equivalent of a 90 would generally find the same success in life as a white child who scored a 97, a difference of about half a standard deviation. The relative take-off in Asian test scores is a recent phenomenon that is driven in part by the children of at least first- and second-children Asian immigrants who are already financially successful compared to the average American. Weberians and other Eurocentric thinkers had pointed to Asians' intellectual inferiority as the reason they couldn't create capitalism, were racially inferior to whites and thus deserved to be conquered. IQ tests simply measure how one has adapted to a society's vision of what intelligence means.

Posted by Reality Man | December 27, 2007 8:14 PM

"The individual diffferences between the races are tiny. 99.99 something % of the human genome are identical across races. Nothing in this 99.99% has changed during the time the races were geologically seperate."
I'm not sure exactly what's intended here, but it seems obvious that phenotypical differences between the races or geographic groupings or whatever have emerged, presumably in the last 50,000 years or so. And the phenotype couldn't have changed except by way of genetic changes. .01% or whatever it is is large enough to account for the differences we all indisputably observe.

"Why do you think IQ is such fertile ground for selection? What leads you to make this assumption."
The assumption would have to be that intelligence aids in attracting mates with reproductively successful traits, and in rearing reproductively successful children to adulthood. (Clark's Farewell to Alms offers an interesting variant of such a hypothesis.) I don't know that smarter men and women have more children who themselves have more children, but it doesn't seem implausible that they do. And I recall the studies that find that intelligence-related genetic traits have observably been under strong selection pressure--have spread through the population with rapidity.

Posted by Mugwump | December 27, 2007 8:24 PM

"The assumption would have to be that intelligence aids in attracting mates with reproductively successful traits, and in rearing reproductively successful children to adulthood."

I don't see why that would lead to differences in IQ between the races. You have to come up with ideas why different populations faced different selection pressures for the kind of thinking tested by IQ tests: pattern recognition, analogies, classification, vocabulary... Sort of scientific thinking, with a big sort of. One of the phrenologists was speculating that the reason why Anasazi jews were supposedly so smart (this is controversial) was they had to live in urban ghettos and were forced to be money lenders. Supposedly they had to be smarter to survive. But why would the smarts they needed to survive be IQ type smarts? You can start supposing that to be a jew in a polish ghetto required you to be constantly aware of the threats around you, constantly evaluating this behavior or that, constantly thinking of new schemes. Of course, there are plenty of studies that show that anxiety makes you stupid, and the Anasazi jews that emmigrated to the US in the early part of the last century suppsoedly had very very low IQs and that generations Charles Murry's were wringing thier hands over the harm admitting these stupid jews would do to society. Or you can speculate as I think it was Rushton who did, that having to deal with the harsh climates of the north made the white man smarter as well as white. But then you ahve to say what was so different about life in the cold north? ANd why did all the early technological advances come from the warmer climates? And how come the Greeks were so smart?

As you can see, once you start trying to immagine the conditions which would cause one group to evolve higher IQ than other groups, you find yourself engagining in the worst kind of speculative thinking, the kind that of thinking that has entertained lots of Michael Crichton and Erik Von Dannekin fans.

In this field I find that there are plenty of "scientists" engaging in this type of speculative thinking and that becomes a factor in assessing the reliability of their work.

Posted by cw | December 27, 2007 9:50 PM

Here is a comment by Jason Malloy of gnxp in the MR comment thread:
"Not controlling for standardized test scores seems like a very basic thing to miss in such a study.

Oh thanks, Mason. They didn't take this into account. Yep, this study appears worthless. When you match by IQ the wage gap disappears."
http://www3.interscience.wiley.com/cgi-bin/abstract/110558275/ABSTRACT?CRETRY=1&SRETRY=0

Malloy has other comments in the thread well worth reading.

Posted by pjgoober | December 27, 2007 9:53 PM

Matt,

So, when are you going to tell us your opinion on the firing of James Watson?

Steve

Posted by Steve Sailer | December 28, 2007 12:51 AM

If we consider measured intelligence to be like a radio signal, the "noise" introduced by historical and cultural factors is overwhelming compared to the "signal" of some mythical "true" measure of varied intelligence across ethnicities. Moreover, we don't even have a very good idea of what it is we're measuring. Furthermore, decoding the human genome has shown us that the leap from gene to observed differences is much more complicated than those who would have us believe that there's just a handful of easily-interpreted "smart genes."

Given all of that, I would say that we know about how much of a role genetics plays in ethnic variation in intelligence as we do about what flavor jam aliens from the other side of the galaxy like to put on their toast. The reality of the world today is such that any genetic variation that may exist (black people could be inherently smarter than whites, or vice versa, or the two could be completely identical) is so obscured by environmental and historical effects that speculating about what that difference may be is pure conjecture.

Posted by Adam Villani | December 28, 2007 1:09 AM

If anybody is interested in a case study of how well Gary Becker's theory of discrimination applies to the most closely studied industry in the world -- professional baseball -- here's my 1996 National Review cover story "How Jackie Robinson Desegregated America:"

http://www.isteve.com/JackieRobinson.htm

"In the liberal world-view, discrimination stems from prejudice, from ignorance of the actual talents of blacks. In organized baseball, the opposite was true. White Major Leaguers freely admitted that many blacks could have taken white players' jobs. Yet, somehow, this enlightened perception failed to make the white pros into ardent integrationists. Meanwhile, a number of owners and managers tried to cheat on their gentlemen's agreement. For example, many historians claim that the Washington Senators quietly broke the color barrier in the late 1930s by playing Cubans dark enough to have been banned as Negroes if they had spoken English.

"There was strikingly little correlation between the rectitude of the man and his urge to integrate baseball. For example, among managers the most creative was the choleric John J. McGraw, a ferocious scrapper who won ten pennants. In 1901 he almost succeeded in smuggling a light-skinned black second baseman onto his team as a full-blooded Cherokee named ``Chief Tokohama.'' During World War II huckster Bill Veeck tried to buy the dreadful Philadelphia Phillies and stock them with Negro Leagues stars. Like all direct challenges, though, this was rebuffed by the autocratic Commissioner of Baseball, Judge Kenesaw Mountain Landis. After the Chicago Black Sox threw the 1919 World Series, the owners had restored faith in the game by appointing Landis, who was famed for his strict moral standards -- one of which was Segregation Forever. After the good Judge went to his reward in late 1944, the owners, hoping to lighten up, picked as Commissioner the Southern politician A. B. ``Happy'' Chandler. When Happy surprisingly indicated that he wouldn't veto black players, Branch Rickey set his plans in motion."

Posted by Steve Sailer | December 28, 2007 1:12 AM

Steve,

You might be interested in the "Guesting Games" thread, about immigration and guest workers.

Posted by pjgoober | December 28, 2007 1:30 AM

Lahn, et al, Science 9 September 2005:
“The gene Microcephalin (MCPH1) regulates brain size and has evolved under strong positive selection in the human evolutionary lineage.”

Balter reporting on the findings:
Lahn and colleagues also found a pronounced pattern in the distribution of the favored alleles in populations around the world: The microcephalin allel, for example, is much more common in Europe, Asia, and the Americas than in sub-Saharan Africa. Using a larger sample from 1184 individuals, the team found this allele in roughly 75% or more of Italians, Russians, and Han Chinese, and in nearly 100% of Colombians. In contrast, the allele had frequencies of less than 10% in the Zime of Cameroon and the San of Namibia, and about 30% of Tanzanian Masaai. The ASPM allele also showed a skewed geographic distribution.

So we know that two alleles arguably (but not certainly, the matter is contentious) brain-related have (a) arrived in the genome as recently as ~40,000 and ~6,000 years ago and (b) are more frequently found in Italian, Chinese, and Columbian populations than among the Zime and Masaai.

“You have to come up with ideas why different populations faced different selection pressures for the kind of thinking tested by IQ tests: pattern recognition, analogies, classification, vocabulary...”

I have to come up also with some explanation why these alleles appeared when they did and spread more rapidly among some populations than others.

In either case, I would assume that the initial appearance an allele in a particular segment of the population can be a random event. Selection pressures I would merely assume aren’t entirely uniform across societies.

Attempting to take a moderate stance, the Gale Encyclopedia of Psychology remarks (Nature-Nurture Controversy, “No one questions the genetic basis of intelligence, but scientists still do not know how intelligence is inherited and what specific aspects of intelligence can be linked to genetic factors.”

If intelligence has a genetic basis, then an allele that raises (or possibly lowers, despite selection for offsetting favorable effects) may appear in one population and not another. This possibility cannot be wished away. But I know of no evidence that the many environmental explanations of racial differences in “IQ” don’t suffice, and I’m sure that there’s much work to be done toward improving the environment here in America.

Posted by Mugwump | December 28, 2007 1:41 AM

Why do I get the feeling that Mugwump is a guy who thought he was the smartest kid in AP Bio in high school and never really took much science beyond that? His posts feel like watching a self-identified communist who has never studied Marxist or any other forms of economic modeling trying to explain why comparative advantage is evil.

Posted by Reality Man | December 28, 2007 6:08 AM

There's nothing wrong with what mugwhump is doing. He's not saying anything (or "speculating") about the genetic inferiority of certain races, at least so far. He's just saying that it's possible. I'm saying that, sure it's possible, but the issue is so complex that no one knows. That allele may affect intelligence in some way, and may effect it enough so that there are noticible differences in the real world, but--from everything I've read--no one knows. Not about the allele or any of the other other evidence. My main point has always been that there are people who say racist things, are connected to white identity groups, who's statistical work is supposedly shoddy, and who are conservative and these people "speculate" freely on this issue, saying all kinds of ridiculous unproven things, proposing all kinds of policies, over and over again like it's their mission in life to prove that black people (it always comes down to black people) are dumber than white people. Why do they do this? Why would anyone choose such a mission? I really really suspect their motivations, but I am not a mind reader. Ask Steve Sailor.

Posted by cw | December 28, 2007 1:47 PM

Hey pjgoober @ 1:30 am.

Can you pick a new name please? Can't you at least add a number after it like "pjgoober2" ? Your post seems like something reasonable I might say, but I was nowhere near a computer anywhere close to 1:30 am today, so I know I didn't just forget I said that.

Posted by pjgoober | December 28, 2007 9:35 PM